Experimental studies of deposition by debris flows: process, characteristics of deposits, and effects of pore-fluid pressure
Abstract
This study examines deposition by experimental debris flows (to 15 m$\sp3$) composed of mixtures of gravel (to 32 mm), sand, and loam using the 95-m-long, 2-m-wide, 31$\sp\circ$ U.S. Geological Survey debris-flow flume. It examines the depositional process, relations between flow kinematics and deposit character, and fluid pressure in debris at, and following, deposition. These data permit evaluation of competing hypotheses regarding debris-flow deposition.Experimental debris flows invariably developed surges; deposits developed abruptly on a 3$\sp\circ$ runout slope as sediment transported by shallow $(<$10 cm deep) surges accumulated. Finer source debris formed thinner deposits. Deposits developed by a combination of forward pushing, vertical accretion, and lateral shunting of previously deposited debris. Unsaturated debris flows deposit sediment dominantly through forward pushing and sourceward horizontal accretion; deposits are lobate, have large aspect ratios $(\geq$0.5), and commonly exhibit arcuate surface ridges. Saturated debris flows progressively deposit sediment primarily through vertical accretion from successive surges; deposits are lobate, but elongate, have small aspect ratios $(\leq$0.3), nearly flat surfaces, and lack prominent surface ridges. Observed progressive accretion is contrary to commonly assumed en masse sedimentation by debris flows.The depositional process is recorded primarily by deposit morphology and surface texture and is not faithfully registered by interior sedimentary texture. Homogeneous internal textures can be interpreted as the result of deposition by a single surge. Individual debris flows as well may leave little distinctive signature in the sedimentary record. Superposed deposits from similar yet separate flows could not be distinguished without the aid of an artificial marker horizon. These results show that methods of estimating flow properties from deposit thickness or from relations between particle size and bed thickness are in error. Relations between sediment composition and deposit thickness are incompatible with deposition by a simple viscoplastic material.Experimental debris flows deposited sediment despite measured basal fluid pressures that were lithostatic to nearly lithostatic. These data refute hypotheses that propose uniform fluid-pressure dissipation as a control on deposition. Modeling and laboratory analyses of gravity-driven consolidation reveal that characteristic pressure-dissipation times in quasistatic debris exceed surge periods and durations typical of debris flows. Numerical simulations of transient changes in fluid-pressure and effective-stress fields in 2-dimensional quasistatic domains reveal that excess fluid pressures remain elevated, and effective stresses depressed, everywhere except adjacent to margins in wide thin bodies for time scales that exceed durations typical of debris flows. Observed deposition, measured fluid pressures, and modeling results suggest that debris-flow deposition is controlled by a balance between a diminishing driving stress and locally increasing resisting stresses along flow margins rather than by a uniform bodywide increase of effective stress.
Collections
- Geology [14]